|Numéro de publication||US7900254 B1|
|Type de publication||Octroi|
|Numéro de demande||US 10/350,183|
|Date de publication||1 mars 2011|
|Date de dépôt||24 janv. 2003|
|Date de priorité||24 janv. 2003|
|État de paiement des frais||Payé|
|Numéro de publication||10350183, 350183, US 7900254 B1, US 7900254B1, US-B1-7900254, US7900254 B1, US7900254B1|
|Inventeurs||Lee Adam Fisher, Gregory Charles Day, Jason Brown|
|Cessionnaire d'origine||Mcafee, Inc.|
|Exporter la citation||BiBTeX, EndNote, RefMan|
|Citations de brevets (13), Référencé par (10), Classifications (6), Événements juridiques (3)|
|Liens externes: USPTO, Cession USPTO, Espacenet|
1. Field of the Invention
This invention relates to the field of data processing systems. More particularly, this invention relates to the identification of malware infected reply messages, such as, for example, reply email messages infected with computer viruses, worms, Trojans etc.
2. Description of the Prior Art
An increasingly common and serious malware threat is that of email propagated computer viruses, worms, Trojans and other items of malware. Using email propagation, such items of malware can rapidly multiply and spread to an extent that causes considerable disruption and economic damage. One type of email propagation involves so called mass mailer viruses. When a computer is infected with such a mass mailer virus, then the mass mailer virus sends itself to some or all of the email addresses in the infected computer's email address book. An outbreak of a mass mailer virus can be identified by observing the email usage characteristics of an email server, such as noting the occurrence of a large number of emails being sent to a large number of different recipients, the occurrence of a large number of emails sharing a common title, a common attachment, or other common features. This type of characteristic behaviour can be identified and anti-malware actions, such as quarantining etc, taken even before a new virus has been fully identified and a proper signature identified and distributed. Existing computer programs which serve to monitor email server behaviour to identify this type of mass mailer virus include Outbreak Manager produced by Network Associates, Inc.
A new type of malware has emerged which propagates by email and has the potential for causing considerable damage and yet does not give rise to characteristic patterns of email traffic that can be proactively detected using the known techniques as mentioned above. These so called reply mailer viruses act on an infected computer by waiting for an email to be received from another computer user and then automatically replying to that specific other computer user with an infected reply email. This infected reply email can reuse the message title of the originating email from that other user and the other user will recognise the sender of the infected reply email as a person known to them. Furthermore, the receipt of a reply email of some sort by the other user will not be unexpected since they have just themselves initiated the email exchange. The result is that the recipient of the infected reply email is likely to consider the infected reply email as genuine and open or deal with it in other ways which cause their computer to become infected.
The known techniques for dealing with mass mailing computer viruses are ineffective against reply mailer computer viruses since the reply is generally made to a single user making the increase in email traffic relatively slight, the email titles can be copied from the originating email messages giving no consistent title that identifies an infected email and there is no sending of a single email message to a large group of recipients which could otherwise be suspicious. Thus, until the specific virus signature for the reply mailer virus has been developed and deployed in the email virus scanning systems, then the known types of email scanners are unable to detect and accordingly provide a defense against reply mailer viruses.
Viewed from one aspect the present invention provides a computer program product for controlling a computer to identify a malware infected message, said computer program product comprising:
reply time detecting code operable to detect a reply message generated in less than a threshold reply time from a reference event associated with a corresponding originating message;
attachment status detecting code operable to detect an attachment status of said reply message matching a trigger attachment status; and
infected message identifying code operable to identify said reply message as a malware infected message if said reply message was generated in less than said threshold reply time and said reply message has an attachment status matching said trigger attachment status.
The present technique recognises that relatively distinctive characteristics of the behaviour of a reply mailer virus and other malware operating upon a similar principle is that the reply message is generated quickly and the reply message has an attachment carrying the infection. Infected behaviour is of the type whereby a reply message is generated very rapidly, for example in a matter of milliseconds after the originating message is sent, in a manner which would not occur as a result of a genuine user opening the originating message, reading the originating message, composing a reply to the originating message and then sending that reply. However, this characteristic cannot be used in itself as there are some known legitimate automated reply generating mechanisms, such as out of office mechanisms, which will rapidly generate a reply message when an originating email is received. However, a further feature of an infected reply message is that it will have an attachment carrying the infection. Also requiring this characteristic helps to avoid false triggers based upon legitimate out of office mechanisms which do not utilise attachments but merely send a simple message in the form of plain text indicating, for example, that the recipient will be unable to read or reply until a certain date.
The present technique uses a combination of a reply message being generated in less than a threshold reply time and an attachment status of the reply message matching a trigger attachment status in order to identify a message as an infected message. This mechanism is able to provide a defense against malware, such as reply mailer viruses, by detecting their characteristic behaviour rather than having to wait until a specific virus detecting signature has been developed and deployed. The higher level of protection against this new type of malware threat is strongly advantageous.
The rapid generation of the reply message which is detected is relative to a reference event associated with the originating message. This reference event could take a variety of different forms, but in preferred embodiments is one of the sending of the originating message or the reading of the originating message. The reading of the originating message may be the reading of the originating message by the mail client program of the receiving user from the mail server database as opposed to the reading of the message by a human user.
The attachment status associated with the reply message could take a variety of different forms. A simple type of attachment status may be the presence or absence of an attachment of any sort. The presence of an attachment of any sort may be taken as a characteristic of the type of malware behaviour for which identification is being sought. Increased selectivity and reduced false alarms can be achieved by seeking to detect attachment status corresponding to the presence of an attachment having a size exceeding a threshold attachment size (this type of malware behaviour requires an amount of coding which is relatively large and thus provides a practical minimum size for an attachment that may carry an infection) or an attachment file type matching a type which can carry computer code to be execute, such as an EXE, COM or VBS file type or the like.
The originating message and the reply message can take a variety of different forms. Current message forms to which the present technique is directly applicable are email messages. However, it is possible that the technique may be applied to other types of message that can carry malware.
It will be appreciated that the threshold reply time varies depending upon the particular characteristics of the system operating the technique, e.g. the processing speed, communication link speed, memory capacity etc. In order to deal with this variability, preferred embodiments of the invention provide a calibration mechanism whereby a calibrating originating message is generated, a calibrating reply message automatically generated and the time between the generation of the calibrating originating message and the generation of the calibrating reply message measured to provide a basis upon which the threshold reply time can be set. The calibrating behaviour is intended to follow the type of behaviour that this sort of malware would show and accordingly be subject to roughly similar processing times giving at least a starting point for the determination of threshold reply times.
These threshold reply times can be calculated for a mail server as a whole, individual users on the mail server or possibly even individual users on a mail server in respect of mail being sent to specific other mail servers or within the same mail server.
It will be appreciated that the malware being identified could take a wide variety of different forms, such as a computer program virus, a computer program worm, a computer program Trojan etc. Other types of malware may also share this type of behaviour.
Viewed from another aspect the present invention also provides a method of identifying a malware infected message, said method comprising the steps of:
detecting a reply message generated in less than a threshold reply time from a reference event associated with a corresponding originating message;
detecting an attachment status of said reply message matching a trigger attachment status; and
identifying said reply message as a malware infected message if said reply message was generated in less than said threshold reply time and said reply message has an attachment status matching said trigger attachment status.
Viewed from a further aspect the invention provides apparatus for identifying a malware infected message, said apparatus comprising:
reply time detecting logic operable to detect a reply message generated in less than a threshold reply time from a reference event associated with a corresponding originating message;
attachment status detecting logic operable to detect an attachment status of said reply message matching a trigger attachment status; and
infected message identifying logic operable to identify said reply message as a malware infected message if said reply message was generated in less than said threshold reply time and said reply message has an attachment status matching said trigger attachment status.
The above, and other objects, features and advantages of this invention will be apparent from the following detailed description of illustrative embodiments which is to be read in connection with the accompanying drawings.
In operation, when a user 10 sends an email message to another user 10 on the same mail server 2, if the other user 10 is infected with a reply mailer virus, then an infected reply email will be automatically generated by the reply mailer virus and received by the user 8 very rapidly and in a consistent amount of time. In a similar way, an email message originating with the user 8 and sent to the user 14 who is a user of a co-operating mail server 4 will also typically show relatively rapid and consistent reply times when the user 14 is infected.
If the originating email message from the user 8 is being sent to the user 16 on a distant mail server 6, then it is possible that the reply times will be relatively long even when the user 16 is infected. However, whilst it may be more difficult to use the present technique to detect infected reply emails from the mail server 6, this may still be done in some circumstances. Furthermore, it is generally the case that a typical user will receive a large number of emails from within their own organisation and from users connected to their own mail server compared with the number received from elsewhere. Thus, detection of reply mailer computer virus characteristics within a mail server or within a group of consistently connected mail servers is highly beneficial and provides a good degree of protection.
At step 28 a check is made for originating emails being sent from an originating user. When such an originating email is sent, then step 30 serves to generate a stored reference event log entry for the originating email indicating the email address of the sender, the email address of the recipient and the time that the originating email was sent. Processing can then return to step 28 to await the next originating email.
If the determination made at step 34 was that the received email has a sender and recipient matching a recipient and sender in the log, then it will be determined to be a reply email. An additional check for a matching title may be made for improved selectivity, although this is not necessarily required and could lead to a loophole that a virus writer could exploit.
If the determination at step 34 was that the received email was not a reply email, then the thread terminates, or in practice returns to step 32 to await the next received email. If the determination at step 34 was that the received email is a reply email, then processing proceeds to step 36. Step 36 compares the elapsed time from the current time at which the reply email has been received with the logged time for the sending of the originating email and determines if this difference is less than a threshold reply time. If this determination is that the elapsed time is greater than the threshold, then the thread again terminates. If the determination is that the elapsed time is less than the threshold, then processing proceeds to step 38 at which a determination of the attachment status of the received reply email is made.
The attachment status can take a variety of different forms. A simple form is merely to check whether or not there is an attachment. An attachment is required by a reply mailer virus to carry the infection (virus code) to a new victim. Accordingly, the absence of an attachment may be taken to indicate that the received reply email is not infected and may be safely passed to the intended recipient without risk. An example of a rapidly generated reply email without an attachment would be an out of office message generated in response to an originating email message. Further examples of attachment status characteristics being matched could be that the received reply email contains an attachment over a predetermined size sufficient to carry a reply mailer virus or that the attachment has a file type suitable for carrying computer program code that can infect a victim computer.
If the determination at step 38 was that the received reply email has an attachment status matching the trigger attachment status, then processing proceeds to step 40 at which infected email actions, such as quarantining, disinfection, deletion, alert message generation etc may be triggered. If the determination at step 38 was that the attachment statement did not match the trigger attachment status, then the tread will be terminated or returned to step 32 without triggering the infected email actions at step 40.
It will be appreciated that the reference events that are logged could alternatively take the form of recording the time at which an email message was read (notified to the email client of the intended recipient). It is normally this reading/notification at an infected user which enables the reply mailer virus computer program at that infected user to obtain the email address of a new victim. Triggering off the reading/notification time rather than the sending time may be advantageous in dealing with situations in which the recipient has been offline (e.g. not yet logged in for the day) and then becomes active and accordingly allows their infected computer to rapidly generate infected reply messages once they have connected to the email server where their email has been waiting for them.
The mail server 62 via which all of these email messages are routed is first to inspect the reply email message 60. It utilises its stored reference event log 64 to determine if the reply email message is a reply or a new email message that should be separately considered. If the reply email message 60 is identified as such, then the mail server 62 references its store of threshold data 66 to determine if the reply email message 60 has been generated in less than a threshold reply time. If the reply email message 60 has been generated in less than the threshold reply time associated with the originating user and the mail server to which the originating email message 60 was sent, then a check is made as to whether or not the reply email message includes an attachment. If the email message does include such an attachment, then the infected message mechanisms are triggered and the reply email message 60 sent to a quarantine store 68 or otherwise processed, such as by disinfection, deletion, triggering generation of an alerting message or the like. If either the reply email message 60 was not generated in less than the threshold reply time or the reply email message 60 did not have an associated attachment, then it is treated as a genuine reply message and returned to the user 54.
Although illustrative embodiments of the invention have been described in detail herein with reference to the accompanying drawings, it is to be understood that the invention is not limited to those precise embodiments, and that various changes and modifications can be effected therein by one skilled in the art without departing from the scope and spirit of the invention as defined by the appended claims.
|Brevet cité||Date de dépôt||Date de publication||Déposant||Titre|
|US5832208 *||5 sept. 1996||3 nov. 1998||Cheyenne Software International Sales Corp.||Anti-virus agent for use with databases and mail servers|
|US6178505||4 mars 1998||23 janv. 2001||Internet Dynamics, Inc.||Secure delivery of information in a network|
|US6266681 *||8 avr. 1997||24 juil. 2001||Network Commerce Inc.||Method and system for inserting code to conditionally incorporate a user interface component in an HTML document|
|US6785732 *||11 sept. 2000||31 août 2004||International Business Machines Corporation||Web server apparatus and method for virus checking|
|US20020049610 *||1 oct. 2001||25 avr. 2002||Gropper Robert L.||Auto update utility for digital address books|
|US20030065941 *||5 sept. 2001||3 avr. 2003||Ballard Clinton L.||Message handling with format translation and key management|
|US20030097409 *||5 oct. 2001||22 mai 2003||Hungchou Tsai||Systems and methods for securing computers|
|US20040107362 *||3 déc. 2002||3 juin 2004||Tekelec||Methods and systems for identifying and mitigating telecommunications network security threats|
|US20040131056 *||19 déc. 2003||8 juil. 2004||Deep Nines, Inc.||Intelligent feedback loop process control system|
|US20040177120 *||7 mars 2003||9 sept. 2004||Kirsch Steven T.||Method for filtering e-mail messages|
|US20050227668 *||8 avr. 2004||13 oct. 2005||Motorola, Inc.||Detection of cloned communication units based on message contents|
|WO2001060019A1||16 janv. 2001||16 août 2001||Secr Defence||Computer security system|
|WO2002013469A2||8 août 2001||14 févr. 2002||Tumbleweed Comm Corp||Recipient-specified automated processing in a secure data file delivery system|
|Brevet citant||Date de dépôt||Date de publication||Déposant||Titre|
|US8234376 *||7 juin 2007||31 juil. 2012||Kabushiki Kaisha Toshiba||Server apparatus and method of preventing denial of service attacks, and computer program product|
|US8984644||28 sept. 2014||17 mars 2015||Securityprofiling, Llc||Anti-vulnerability system, method, and computer program product|
|US9100431||28 sept. 2014||4 août 2015||Securityprofiling, Llc||Computer program product and apparatus for multi-path remediation|
|US9117069||21 déc. 2013||25 août 2015||Securityprofiling, Llc||Real-time vulnerability monitoring|
|US9118708||28 sept. 2014||25 août 2015||Securityprofiling, Llc||Multi-path remediation|
|US9118709||28 sept. 2014||25 août 2015||Securityprofiling, Llc||Anti-vulnerability system, method, and computer program product|
|US9118710||29 sept. 2014||25 août 2015||Securityprofiling, Llc||System, method, and computer program product for reporting an occurrence in different manners|
|US9118711||29 sept. 2014||25 août 2015||Securityprofiling, Llc||Anti-vulnerability system, method, and computer program product|
|US9225686||16 mars 2015||29 déc. 2015||Securityprofiling, Llc||Anti-vulnerability system, method, and computer program product|
|US20080104182 *||7 juin 2007||1 mai 2008||Kabushiki Kaisha Toshiba||Server apparatus and method of preventing denial of service attacks, and computer program product|
|Classification aux États-Unis||726/22|
|Classification coopérative||G06F21/566, G06Q10/107, H04L51/08, H04L51/12|
|24 janv. 2003||AS||Assignment|
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:FISHER, LEE ADAM;DAY, GREGORY CHARLES;BROWN, JASON;REEL/FRAME:013698/0744
Effective date: 20030120
Owner name: NETWORKS ASSOCIATES TECHNOLOGY, INC., CALIFORNIA
|1 août 2005||AS||Assignment|
Owner name: MCAFEE, INC., CALIFORNIA
Effective date: 20041119
Free format text: MERGER;ASSIGNOR:NETWORKS ASSOCIATES TECHNOLOGY, INC.;REEL/FRAME:016593/0812
|20 août 2014||FPAY||Fee payment|
Year of fee payment: 4